TY  - RPRT
TI  - A Reproducible Optimisation Protocol for Calibrating Prompt-Based Large Language Model Workflows in Evidence Synthesis
AU  - Teo Susnjak
PY  - 2026
UR  - https://arxiv.org/abs/2605.06937
ID  - 2605.06937
ER  -